# Lag

Lag shifts your features on a partition index, creating a lookback feature offset by an amount. Lag supports generating multiple lags in one transform by generating each unique combination of columns and amounts from your inputs.

## Parameters

| Name      | Type         | Description                                                                                                                                         | Is Optional |
| --------- | ------------ | --------------------------------------------------------------------------------------------------------------------------------------------------- | ----------- |
| columns   | column\_list | names of column(s) you want to lag                                                                                                                  |             |
| amounts   | int\_list    | Magnitude of amounts you want to use for the lag. Positive values result in a historical offset; negative amounts result in forward-looking offset. |             |
| partition | column\_list | name of column(s) to partition by for the lag                                                                                                       | True        |
| order\_by | column\_list | name of column(s) to order by in the final data set                                                                                                 | True        |

## Example

```python
ds = rasgo.get.dataset(id)

ds2 = ds.lag(columns=['OPEN', 'CLOSE'], amounts=[1,2,3,7], order_by=['DATE, 'TICKER'], partition=['TICKER'])
ds2.preview()
```

## Source Code

{% embed url="<https://github.com/rasgointelligence/RasgoTransforms/blob/main/rasgotransforms/rasgotransforms/transforms/lag/lag.sql>" %}